Noise level at base ISO pushed vs. higher ISO

[Jan Brittenson]

Hmm, I think the C1 exposure slider stretches the histogram, clipping the top end - right? This is a little different from using a curves panel and raising the middle of the curve, or a little below middle (maybe 1/3 up the scale). The former is useful when the high end of the histogram is empty and the image is bunched up at the lower end. The latter when most of the image is bunched up at the lower end but it's still desirable to retain what's at the high end (and what would normally have been clipped with more exposure).

 

[thomas]

Hmm, I think the C1 exposure slider stretches the histogram, clipping the top end - right? This is a little different from using a curves panel and raising the middle of the curve

for this topic a curve is not appropriate. We are talking about a file being 2 stops underexposed. So you have to recover 2 stops in the shadows (actually 2 stops in the entire image) - you can't recover shadows by raising the middle of the curve.
Too, this was not about "what's the best way to recover an underexposed image". It was simply about the comparison of hardware based ISO-push (digiback) vs. software ISO-push.
Too, a capture underexposed by 1 stop but pushed with "exposure" in the software and a capture taken with 1 stop higher ISO look the same (literally), i.e. the dispersion of luminance over the tonal range is the same (the same goes for 2 stops and 2.5 stops).
Okay, the image with the higher ISO may show little less DR (maybe around 0.1EV or so)... but that's just a little difference and not relevant here.

 

[Professional]

Sorry i didn't do any test yet because i was lazy, tired and a bit busy, but last question, can i use artificial lights [studio strobes, speedlights,...] for this test or i should keep natural light available [room light, outdoor, windows,...]?

 

[Guy Mancuso]

Tungsten is the worst source for noise. I always use tungsten light when testing for noise.

 

[Professional]

Tungsten is the worst source for noise. I always use tungsten light when testing for noise.

I agree it is the worst light for noise, ok then, i will use normal tungsten room light for the test even we have florescent as well], so in this case i should use tungsten WB.

 

[thomas]

can i use artificial lights [studio strobes, speedlights,...] for this test or i should keep natural light available?

eihter way. It's not about the noise level. It's only about the relation of base ISO pushed and shotting at the respective ISO setting.

 

[Professional]

Hi People, i did the test but there is something wrong with the results, so that i don't know if i did the test correct or not.

 

[Professional]

ISO 50


ISO 50 with +1 exposure


ISO 50 with +2 exposure


ISO 100


ISO 200

 

[thomas]

Hi People, i did the test but there is something wrong with the results, so that i don't know if i did the test correct or not.

The shots are fine. Now you have to push the 1 stop underexposed ISO50 shot in the RAW software by 1 stop and the 2 stops underexposued ISO50 by 2 stops.... so taht all the catpure have (about) the same luminance level.

 

[Professional]

The shots are fine. Now you have to push the 1 stop underexposed ISO50 shot in the RAW software by 1 stop and the 2 stops underexposued ISO50 by 2 stops.... so taht all the catpure have (about) the same luminance level.

Sorry that i didn't clarify my post above, in fact #2 and #3 posted above are pushed already as written in the ACR [#2 with +1 or 1 stop pushed in exposure, #3 is pushed with +2 exposure, is +1 in exposure means 1 stop?]

 

[Jan Brittenson]

you can't recover shadows by raising the middle of the curve.

Of course you can. Raise any point on the curve and you increase contrast below that point while decreasing above it. Increasing contrast brings out detail, hence raising any point on a curve increases shadow detail by virtue of increased contrast.

 

[thomas]

Of course you can. Raise any point on the curve and you increase contrast below that point while decreasing above it. Increasing contrast brings out detail, hence raising any point on a curve increases shadow detail by virtue of increased contrast.

basically you are right.
But for the comparision of a pushed underexposed shot and a shot at the respective higher ISO setting "exposure" to +1 (resp. +2) preserves the dispersion of luminance of both the captures. Due to the "magnetic" (?) design of the curves in C1 you'll have to make multiple points in the curve (and one of them has to be at the very low end of the curve) so I think you'll have a hard time to recover the same dispersion of light all over the histogram. With "exposure" it's one click here.

@Professional: why don't you take the Hasselblad software? I don't think ACR will do it here.
The 1 stop underexposed ISO50 shot pushed by the exposure value "1" should look like the ISO100 shot and the 2 stops underexposed ISO50 shot pushed by the EV "2" should look like the ISO200 shot.
If not that's probably a software issue.

 

[Professional]

basically you are right.
But for the comparision of a pushed underexposed shot and a shot at the respective higher ISO setting "exposure" to +1 (resp. +2) preserves the dispersion of luminance of both the captures. Due to the "magnetic" (?) design of the curves in C1 you'll have to make multiple points in the curve (and one of them has to be at the very low end of the curve) so I think you'll have a hard time to recover the same dispersion of light all over the histogram. With "exposure" it's one click here.

@Professional: why don't you take the Hasselblad software? I don't think ACR will do it here.
The 1 stop underexposed ISO50 shot pushed by the exposure value "1" should look like the ISO100 shot and the 2 stops underexposed ISO50 shot pushed by the EV "2" should look like the ISO200 shot.
If not that's probably a software issue.

I will check the files again and try the Hasselblad software if i will see a difference, i tried also to push the exposure from Photoshop after open not by ACR and still same problem, so is Photoshop also not reliable software as well, why will that Hasselblad software will do better job in exposure here?!!!

 

[Jack]

You have to push exposure in the raw processor -- I would use Hassy's proprietary software, not ACR/LR since they treat "exposure" differently depending on manufacturer...